Dalton Transactions View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

PAPER

Cite this: Dalton Trans., 2015, 44, 10795

View Journal | View Issue

Core-level photoemission spectra of Mo0.3Cu0.7Sr2ErCu2Oy, a superconducting perovskite derivative. Unconventional structure–property relationships† Sourav Marik,a,b Christine Labrugere,b O. Toulemonde,b Emilio Morána and M. A. Alario-Franco*a Detailed studies of the electronic states for Mo0.3Cu0.7Sr2ErCu2Oy samples with different oxygen contents are presented here. The influence of oxygenation on the electronic states for the Mo0.3Cu0.7Sr2ErCu2Oy system from the semiconducting to the superconducting state has been investigated by means of X-ray photoelectron spectroscopy (XPS). The XPS studies show that Mo is in a mixed MoV and MoVI oxidation state and MoV is predominant over the MoVI in the as-prepared (AP) sample. Yet annealing under an oxygen atmosphere enhances the MoVI state. At the same time, a reduction in the copper species is observed. In the Cu 2p spectra, a larger energy separation between the satellite and main peaks (ES–EM) and a lower intensity ratio (IS/IM) are found to correlate with higher values of the superconducting transition temperature (TC). Analysis of these spectra within the Configuration Interaction (CI) model suggests that higher values of TC are related to lower values of the O 2p–Cu 3d charge transfer energy. The change in the Sr 3d and O 1s core level spectra correlates with the oxygen insertion in the (Mo/Cu)O1+δ chain site, after oxygenation. The hole concentration (Ph) in the copper plane has been obtained using the room temperature thermoelectric power (TEP) value; this shows an increasing tendency with increas-

Received 1st February 2015, Accepted 19th February 2015

ing TC, after oxygenation. From these experimental results, one observes that TC increases with decreas-

DOI: 10.1039/c5dt00459d

ing charge transfer energy. This is, indeed, opposite to the accepted views and occurs in parallel with the shortening of the apical copper–oxygen distance (Cu2–O2) and the increasing of the CuO2 plane buck-

www.rsc.org/dalton

ling angle.

1.

Introduction

The perovskite structure, the most abundant structure in the Earth’s mantle, of which about 93% is believed to be the orthorhombic (Mg,Fe)SiO3,1 appropriately called Bridgmanite,2 is also an extremely important structure on the Earth’s surface, given its great relevance to science and technology. Concerning its relevance, and this themed issue of Dalton Transactions is obvious proof of it, one only has to recall the considerable number of materials, and material types, that have this beautifully simple, structurally flexible and compositionally adaptable chemical species, and the almost

a Dpto. Química Inorgánica, Facultad de CC.Químicas, Universidad Complutense de Madrid, 28040-Madrid, Spain. E-mail: [email protected] b CNRS, Université de Bordeaux, ICMCB, 87 avenue du Dr A. Schweitzer, Pessac, F-33608, France † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c5dt00459d

This journal is © The Royal Society of Chemistry 2015

unlimited number of useful solids that share, rarely, the aristotype and more often than not, one of its multiple variants.3 Among perovskite materials, the celebrated high temperature superconducting cuprates, given birth to by the apparently modest but, in the end, revolutionary paper on La2−xBaxCuO4 and its congeners by Bednorz and Muller,4 constitute the most significant family of superconductors discovered in the last 28 years;5 but perovskites also occupy a prominent position in the different variants of multiferroicity, giant magnetoresistance, catalysis and various types of sensors,6 to name but a few. It is widely accepted that high temperature superconductivity in cuprates7 arises from doping of a Mott–Hubbard insulator.8 Understanding metal–insulator transitions of doped Mott–Hubbard insulators is one of the outstanding challenges occupying modern solid state physics. In particular, in the doped cuprates, there exists a complex interplay between lattice, charge, and spin degrees of freedom. Various phases between the antiferromagnetic charge-transfer insulator and the paramagnetic Fermi-liquid like state have been the subject

Dalton Trans., 2015, 44, 10795–10805 | 10795

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Paper

Dalton Transactions

of discussion. The prominent state of matter in this context is superconductivity. Materials with the 1212 type-structure9 (MA2RECu2O8−δ or M-1212, where M, A and RE are commonly a transition metal element, an alkaline earth metal and a rare earth ion, respectively) have drawn special attention as they are close to the YBa2Cu3Oy 10 structure (variously represented as YBCO, CuBa2YCu2Oy (CuSr2YCu2Oy, YSCO) or Cu-1212) having copper in pyramidal coordination with the ligand oxygen. From the Cu-1212-type structures, Mo-stabilized CuSr2RECu2Oy (RE = rare earth elements) single phase compounds can be prepared at ambient pressure until reaching the composition Mo0.3Cu0.7Sr2RECu2Oy ((Mo,Cu)-1212).11 This is, indeed, an exciting system to study superconductivity adjacent to a metalinsulator boundary.11–13 Previous work on Mo-stabilized CuSr2RECu2Oy phases showed that these compounds display a superconducting transition centring around 30 K associated with an oxidation reaction from a semiconducting state, with the exception of the bigger RE = La, Pr and Nd compounds, which are not superconducting.14 However, the superconducting transition temperatures of Mo0.3Cu0.7Sr2RECu2Oy materials can be almost as high as that of YBCO when they are oxygenated under high pressure.13,15–17 Interestingly, in these compounds, Mo partially adopts the MoV oxidation state, which induces magnetic interactions within the charge reservoir layer.11 As a result, the as-prepared molybdo-cuprate compounds show interesting physical properties.11–13,18,19 Previous crystal structure studies11–13,16,20–22 show that all the Mo0.3Cu0.7Sr2RECu2Oy compounds adopt a tetragonal structure with the P4/mmm space group. It is noteworthy that in contrast to the two-orbital model of Sakakibara et al.23 which shows that TC increases with the increasing apical copper–oxygen (Cu2–O2) distance, we have found that, in this system, the apical Cu2–O2 distance decreases after oxygenation with increasing TC for all the 1212type Mo0.3Cu0.7Sr2RECu2Oy compounds.12,13,16 Moreover, the structural refinement of the as-prepared (non-superconducting) and oxidized (superconducting) phases for the first four members of the homologous series (Cu0.75Mo0.25)Sr2(Ce,Y)sCu2O5+2s+δ by Grigoraviciute et al.17 showed that, besides the 1212 compound, the 1222 and 1232 molybdo-cuprate compounds also show the trend of shortening apical oxygen dis-

tances after oxygenation. At the same time, in these compounds,12,13,16 the out-of-plane buckling angle24 increases with oxygenation in disagreement with the common belief that maximum TC’s are achieved when perfectly flat CuO2 planes are present, as exemplified by mercury cuprate.25 High-temperature superconductors have been widely studied by photoemission spectroscopy;26 X-ray photoelectron spectroscopy (XPS) is a powerful tool for studying the doping of the 3d element in the Cu site and can study the electronic structure of the system. Core-level measurements provide information on the electronic states of the compounds, such as oxidation states and doping-induced chemical potential shifts and, in the case of the highly correlated Cu cations, charge-transfer mechanisms and multiplet splitting. The lack of studies on the electronic structure of this interesting molybdo-cuprate family prompted us to investigate the electronic states of 1212-type molybdo-cuprate materials having different oxygen contents and to relate them to superconductivity, in particular to TC. We present and discuss in this paper a detailed XPS study of various Mo0.3Cu0.7Sr2ErCu2Oy, a superstructure of perovskite, having different oxygen contents and the influence of that on the materials electronic properties, in particular its critical temperature. The evolution of the physical properties with oxygen doping from semiconducting to superconducting, and its consequences on the electronic states, has been investigated and is discussed.

2. Experimental details The details of the synthesis procedure have been presented elsewhere.11–13 The sample compositions, annealing (oxygenation process) conditions, and their naming details are presented in Table 1. The sample purity was checked by powder X-ray diffraction. The oxygen contents were obtained by refining the neutron powder diffraction (NPD) data13 and thermogravimetric analysis (TGA, data taken under reducing conditions (5% H2 + 95% N2). Samples were heated to 973 K, then isothermally maintained for 15 h and cooled down to room temperature. The oxygen content was calculated considering that the final pro-

Table 1 Sample annealing (oxygenation process) conditions, their naming details, superconducting transition temperatures (TC), calculated oxygen contents, TEP values at 290 K and the calculated hole concentrations (Ph) for the different Mo0.3Cu0.7Sr2ErCu2Oy samples

Preparation conditions 1273 K for 48 h in air AP sample annealed under O2 flow at 873 K for 48 h OA sample annealed under O2 flow at 673 K for 24 h AP sample annealed at 673 K for 48 hour under high oxygen pressure (pressure = 100 bar) Oxygenation of the AP sample in belt press type apparatus at 5 GPa and 773 K for 30 min in the presence of 33 mol% KClO3 Oxygenation of the AP sample in belt press type apparatus at 8 GPa and 773 K for 30 min in the presence of 33 mol% KClO3

10796 | Dalton Trans., 2015, 44, 10795–10805

Sample name

TC (K)

Oxygen content (y)

S (290 K) in µV K−1

Ph

AP OA OA1 OAHP

NON SC 32 K 25 K 45 K

7.34 (NPD) 7.45 (NPD) 7.36 (TGA) 7.46 (TGA)

117.5 29.6 32.8 24.5

0.035 0.092 0.089 0.097

HPO

84 K

7.55 (NPD)

—

—

HPO_80 kbar

89 K

7.58 (TGA)

—

—

This journal is © The Royal Society of Chemistry 2015

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Dalton Transactions

ducts were Er2O3, SrO, SrMoO4 and Cu as observed from the X-ray diffraction data. Physical property measurements were carried out by means of magnetic susceptibility, resistivity and thermoelectric power measurement (TEP) techniques. The resistivity and TEP measurements were carried out in the absence of an applied magnetic field. Magnetic susceptibility measurements were carried out under a 10 Oe magnetic field. The X-ray photoelectron spectroscopy (XPS) measurements were carried out on a Thermo-VG Scientific ESCALAB 220 iXL spectrometer, equipped with a monochromatic Al Kα X-ray source (1486.6 eV). The samples were scraped under ultra-high vacuum (UHV; 10−9 mbar) at room temperature with a stainless steel blade until strong attenuation of the C ls signal was achieved. The spectra were collected with 20 eV constant pass energy. The collected XPS data are fitted with a combination of Gaussian–Lorentzian line shapes, after correction for the Shirley background, by using the AVANTAGE software from Thermofisher Scientific.

3. Results

Paper

methods, for all the samples, are compared in Table 1. Fig. 2a shows the change of TC with oxygen contents for five different Mo0.3Cu0.7Sr2ErCu2Oy samples. It clearly shows that TC increases with increasing oxygen content in the structure. The materials oxygenated using high pressure show higher superconducting transition temperatures (TC = 84 K and 89 K for HPO and HPO_80 kbar sample, respectively). The hole concentrations (Ph) were calculated from the 290 K TEP (S (290 K)) data using the universal formula.27 S ð290 KÞ ¼ 372 expð32:4 P h Þ for 0:00 , P h , 0:05; S ð290 KÞ ¼ 992 expð38:1 P h Þ for 0:05 , P h , 0:155 S ð290 KÞ ¼ 139 P h þ 24:2 for P h > 0:155:

ð1Þ

The TEP measurements (Fig. 2 in the ESI†) show that the hole concentration increases with increasing oxygen content (Fig. 2b), as does the TC. At the same time, the detailed structural studies12,13 showed that the variation in the superconducting transition temperature is related to short-range oxygen ordering in the chain copper sites ((Mo/Cu)O1+δ) after oxygenation. This indeed increases the effectiveness in the charge transfer to the superconducting layer (CuO2), resulting in superconductivity.

Physical properties, oxygen content and hole concentration Rietveld refinements of the XRD and NPD (for AP, OA and HPO) data of the samples indicate that all the compounds can be isolated as single phases, crystallizing in a tetragonal structure with the space group P4/mmm. The detailed crystal structure using the combination of XRD, NPD and electron diffraction patterns has been discussed elsewhere.13 The crystal structure of Mo0.3Cu0.7Sr2ErCu2Oy along with the schematic diagram is shown in Fig. 1. Magnetic susceptibility and resistivity measurements are shown in the ESI Fig. 1.† The AP sample is non-superconducting; however, an oxygenation process induces superconductivity in the material. The superconducting transition temperatures and calculated oxygen contents using different

Fig. 1 Schematic diagram of the (Mo,Cu)-1212Er structure (Mo0.3Cu0.7Sr2ErCu2Oy, tetragonal, P4/mmm) along with the crystal structure as suggested by NPD refinement (see ref. 6 and 7). Two different chain oxygen ions are highlighted in black and deep green; other oxygen ions are shown as red spheres. The detailed structural analysis and oxygen positions are described in ref. 6.

This journal is © The Royal Society of Chemistry 2015

XPS core-level measurements To get a detailed view of the electronic states and the influence of oxygenation on them, notwithstanding that we used a surface probe technique, we have studied the XPS core level spectra for Mo0.3Cu0.7Sr2ErCu2Oy samples with different oxygen contents. However, we were unable to perform XPS measurements (with scraping) on the HPO_80 kbar sample due to the small amount of sample able to be prepared at such high pressure. The XPS core level spectra for the other Mo0.3Cu0.7Sr2ErCu2Oy samples are provided. C 1s core level spectra. Two intense peaks of C 1s core electrons at 285 eV and at 289 eV are observed before scraping (Fig. 3 in the ESI†). This indicates unambiguously two different states of carbon.28 The peak at 285 eV corresponds to the normal contamination of the surface by carbon while the second one, at 289 eV, is due to the formation of carbonates on the surface. The samples were then scraped in a vacuum at room temperature with a stainless steel blade. A clean surface can, generally, be obtained after a couple of mild scraping over the surface. The effect of cleaning is evident in the strong attenuation of the C ls signal. Sr 3d core level spectra. Fig. 3a compares the Sr 3d core level spectra for the Mo0.3Cu0.7Sr2ErCu2Oy samples. All the samples show a doublet corresponding to the transitions from Sr 3d3/2 (134 eV) and Sr 3d5/2 (132.4 eV) separated by 1.6 eV.29 After oxygenation, they clearly show changes in the core level spectra. The intensity of the Sr 3d3/2 spectra increases after oxygenation. Eventually, for the HPO sample having the highest oxygen content (y = 7.55) among all our samples studied using XPS, Sr 3d3/2 shows the most intense peak among all the samples. Also, in the same process, the Sr 3d3/2 spectra are shifted to the lower energy side, with the oxygenation.

Dalton Trans., 2015, 44, 10795–10805 | 10797

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Paper

Dalton Transactions

Fig. 2 (a) Variation of TC with oxygen contents for the different Mo0.3Cu0.7Sr2ErCu2Oy samples. (b) Variation of TC with the hole concentration (Ph), calculated using 290 K TEP data. It shows that the hole concentration increases with increasing oxygen content in the structure, as does the TC.

Fig. 3 (a) Core level XPS spectra for the Sr 3d energy region for the Mo0.3Cu0.7Sr2ErCu2Oy samples. The intensity of the Sr 3d3/2 spectra increases and shifts towards the lower binding energy side after oxygenation. (b) Fitted core level XPS spectra for the Sr 3d energy region for the same. The spectra have been deconvoluted using two doublets. The lower binding energy (LBE) doublet represents the Sr atom in the SrO layer between the copper chain and the CuO2 plane and the higher binding energy (HBE) doublet represents the Sr atom in the rare earth (Er) site, sandwiched between two CuO2 layers.

10798 | Dalton Trans., 2015, 44, 10795–10805

It is worth recalling that the Sr atom (Sr–O) resides between the (Cu/Mo)O1+δ chain and the CuO2 plane in the (Mo,Cu)1212 structure (Fig. 1). The changes in the Sr 3d core level spectra after oxygenation have two possible explanations. (I) Any changes, structural or compositional, in either the copper chain site ((Cu/Mo)O1+δ) or the copper plane should also be reflected in the SrO layer. It is likely that due to the presence of the extra oxygen in the copper chain site ((Cu/Mo) O1+δ), the Sr 3d core level spectra will show changes after oxygen annealing. (II) Another reason, not incompatible with the previous one, attributes the changes in the Sr 3d core level spectra to Sr/ Er disorder. In the Bi-1222 (Bi2Sr2CaCu2O8)29 materials, where Sr/Ca disorder occurs easily, the Sr 3d spectra were deconvoluted into two doublets. The lower binding energy (LBE) doublet represents the Sr atom in the SrO layer between the copper chain and the CuO2 plane, and the higher binding energy (HBE) doublet represents the Sr atom in the rare earth (or Ca site for the Bi-1222 material) sandwiched between two CuO2 layers. Similarly, we have deconvoluted the Sr 3d spectra of all the Mo0.3Cu0.7Sr2ErCu2Oy samples with two doublets. Fig. 3b shows the deconvoluted spectra, and the fitting parameters are shown in Table 2. The deconvolution of the Sr 3d core level spectra shows that the intensity of the HBE doublet increases with oxygenation. This, in fact, points out to the possibility of intermixing between Sr and Er. However, our structural analysis using the joint RT NPD/RT XRD refinement does not show any indication of Sr/Er intermixing.12,13 It then appears that the changes in the Sr spectra after oxygenation are related to the variations in structure and composition produced in the (Cu/Mo)O1+δ chain layers. Er 4d core level spectra. The core level Er 4d spectra for all the Mo0.3Cu0.7Sr2ErCu2Oy samples are shown in Fig. 4. They show a peak at 168.2 eV, which is similar to the binding energy observed in pure erbium oxide.30 No significant change in the

This journal is © The Royal Society of Chemistry 2015

View Article Online

Dalton Transactions

Paper

Table 2 Parameters and summary of fitting of the Sr 3d XPS core level spectra for Mo0.3Cu0.7Sr2ErCu2Oy samples

Sample name

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

AP

OA

OA1

OAHP

HPO

Core-level spectra

B. E (eV)

(LBE) Sr 3d5/2 3d3/2 (HBE) Sr 3d5/2 3d3/2 (LBE) Sr 3d5/2 3d3/2 (HBE) Sr 3d5/2 3d3/2 (LBE) Sr 3d5/2 3d3/2 (HBE) Sr 3d5/2 3d3/2 (LBE) Sr 3d5/2 3d3/2 (HBE) Sr 3d5/2 3d3/2 (LBE) Sr 3d5/2 3d3/2 (HBE) Sr 3d5/2 3d3/2

132.3 134.1 133.4 135.2 132.2 134.0 133.4 135.2 132.0 133.8 133.3 135.1 132.0 133.8 133.3 135.1 132.4 134.2 133.5 135.3

FWHM

I (LBE)/ I (HBE)

1.4

1.00/0.54

1.6 1.5

1.00/0.50

1.6 1.5

0.99/1.00

1.6 1.4

0.99/1.00

1.5 1.5

0.91/1.00

1.5

Fig. 4 Core level XPS spectra of the Er 4d energy region for the Mo0.3Cu0.7Sr2ErCu2Oy samples. No significant change in the Er 4d spectra has been observed after oxygenation. The peak position is found to be the same in all cases.

Er 4d spectra has been observed after oxygenation. Moreover, the peak position is found to be the same in all cases. This, indeed, corroborates the absence of Sr/Er intermixing after oxygenation in the present samples. Mo 3d core level spectra. Fig. 5 compares the XPS spectra of the Mo 3d core levels for the different Mo0.3Cu0.7Sr2ErCu2Oy

This journal is © The Royal Society of Chemistry 2015

Fig. 5 Core level XPS spectra of the Mo 3d energy region for the Mo0.3Cu0.7Sr2ErCu2Oy samples. Both peak broadening and peak shifting are observed for the oxygenated samples compared to the AP one; this indicates the oxidation of Mo with increasing oxygen content in the structure.

samples. All the spectra exhibit a two peak structure, typical of 3d core levels with spin orbit splitting. The Mo 3d5/2 and 3d3/2 peaks for the AP sample are at 232.1 eV and 235.2 eV respectively. We can observe both peak broadening and peak shifting for the oxygenated samples compared to the AP ones. The difference between the 3d5/2 and 3d3/2 peaks is 3.1 eV for all samples. The intensity ratios of the two components are close to 1.5 in all cases, which is expected from the multiplicity of the spin orbit split feature. However, to get an idea of the oxidation state of molybdenum and the cause of peak broadening and peak shifting, we have continued the fitting on the Mo 3d core level spectra. A spin–orbit splitting of 3.1 eV31 and a 3d5/2 : 3d3/2 intensity ratio of 3 : 2 have been used for the fitting, after appropriate correction of the background with a single Shirley (Gaussianstep) function. The fitted spectra are provided in the ESI Fig. 4† and the details of the fitting parameters are shown in Table 3. The fitting of the XPS spectra of Mo 3d core levels allows us to confirm the large content of MoV in the AP sample. MoV and MoVI are identified by the energy shifts of their core levels, assigning the doublet at 232 eV and 235.1 eV (3d5/2 and 3d3/2) to MoV and the one at 232.9 eV and 236.0 eV (3d5/2 and 3d3/2) to MoVI, as proposed before.11 Annealing under an oxygen atmosphere increases the amount of MoVI. For the HPO sample having the highest oxygen content (y = 7.55), we have found the highest amount of MoVI (40%, while in the AP sample MoVI = 24% and y = 7.34; see Table 3). Cu 2p core level spectra. The Cu 2p3/2 spectra for all the samples are shown in Fig. 6. The Cu 2p3/2 spectra for all the samples are typical of CuII compounds. The main peak near 933 eV is attributed to the well-screened 2p53d10L final states

Dalton Trans., 2015, 44, 10795–10805 | 10799

View Article Online

Paper

Dalton Transactions

Table 3 Parameters and summary of fitting of the Mo 3d XPS core level spectra for Mo0.3Cu0.7Sr2ErCu2Oy samples

Sample name with TC in K

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

AP (0 K)

OA (32 K)

OA1 (25 K)

OAHP (45 K)

HPO (84 K)

Core-level spectra V

Mo 3d5/2 3d3/2 MoVI 3d5/2 3d3/2 MoV 3d5/2 3d3/2 MoVI 3d5/2 3d3/2 MoV 3d5/2 3d3/2 MoVI 3d5/2 3d3/2 MoV 3d5/2 3d3/2 MoVI 3d5/2 3d3/2 MoV 3d5/2 3d3/2 MoVI 3d5/2 3d3/2

B. E (eV) 232.0 235.1 232.9 236.0 231.9 235.0 233.0 236.1 232.0 235.1 233.1 236.2 231.9 235.0 233.0 236.1 231.9 235.0 233.0 236.1

MoV/MoVI

Average oxidation state

y

0.76/0.24

5.24

7.34

0.71/0.29

5.29

7.45

0.70/0.30

5.30

7.36

0.69/0.31

5.31

7.46

0.60/0.40

5.40

7.55

medium in an extended configuration-interaction (CI) model34,37,38 is a superposition of the |3d 9> and |3d10L̲> electronic states. The interaction is described by the hybridization integral T = and the metal to ligand (Cu 3d→ O 2p) charge transfer energy Δ = − is a measure of the bonding–antibonding separation in the ground state. However, as shown in Fig. 6, one can see a clear shoulder (shown by an arrow) and broadening of the Cu 2p3/2 main peak at the higher binding energy side for the AP sample, which is a characteristic feature of CuIII.39 Also, the main peak is shifted towards the lower energy side after oxygenation. We have fitted the Cu 2p3/2 spectra for all the samples. The fitting parameters, the summary of the fitting and the average oxidation state of copper are shown in Table 4. Quite interestingly, in parallel to the oxidation of the molybdenum species a reduction of the copper is observed in the XPS spectra of the Cu 2p core levels. The average oxidation state of Cu goes from 2.19 in the AP sample to 2.09 for the HPO sample. Beside the peak shift, the satellite peak intensity is found to decrease after oxygenation. The satellite peak to main peak intensity ratios (IS/IM, Table 4) and their energy separation (ES– EM) are calculated using a single Gaussian fitting in the satellite and main peak, respectively. The IS/IM ratio shows the lowest value for the HPO sample, having the highest oxygen content (y = 7.55). According to the Configuration Interaction (CI) model, the change in the ratio between the satellite peak and the main peak intensity (IS/IM) is associated with a charge transfer energy (Δ) from the O 2p to the Cu 3d state.33–35 This ratio goes in parallel with the charge transfer energy: it decreases when the charge transfer energy decreases. In a simple configuration interaction model, IS/IM and ES–EM, are related to the parameters Δ, T (hybridization integral), and U (on-site Coulomb interactions between Cu 2p and 3d holes): ES  EM ¼ ððΔ  UÞ2 þ 4T 2 Þ1=2 I S =I M ¼ tan2 ðφ  θÞ; where tanð2θÞ ¼ 2T=Δ and tanð2φÞ ¼ 2T=ðΔ  UÞ

ð2Þ

The charge transfer energy (Δ) is estimated by considering T = 2.2 eV, and U = 7.8 eV, as described previously.37 Fig. 7 Fig. 6 Core level XPS spectra of the Cu 2p3/2 energy region for the Mo0.3Cu0.7Sr2ErCu2Oy samples. The CuIII characteristic shoulder is shown by an arrow.

resulting from ligand-to-metal (O 2p→Cu 3d) charge transfer, and the satellite at higher binding energy corresponds to a multiplet of 2p53d 9L states, where bold character denotes a hole and L denotes the oxygen ligand.32–36 In a simplified description of the Cu 2p3/2 spectrum, to maximize the energy gain, a doped hole first extends out as a symmetric coherent superposition of the four O 2p hole states surrounding the CuII ion and then hybridizes with the central Cu dx2−y2 hole to form a Zhang–Rice singlet.36 The ground state of the neutral CuO4 cluster integrated in an effective

10800 | Dalton Trans., 2015, 44, 10795–10805

Table 4 Parameters and summary of fitting of the Cu 2p XPS core level spectra for Mo0.3Cu0.7Sr2ErCu2Oy samples

Sample name with TC in K AP (0 K) OA (32 K) OA1 (25 K) OAHP (45 K) HPO (84 K)

Core-level spectra

B. E (eV)

FWHM

CuII : CuIII

IS/IM

y

CuII 2p3/2 CuIII 2p3/2 CuII 2p3/2 CuIII 2p3/2 CuII 2p3/2 CuIII 2p3/2 CuII 2p3/2 CuIII 2p3/2 CuII 2p3/2 CuIII 2p3/2

933.1 934.8 933.2 935.2 933.1 935.0 933.0 934.8 933.1 934.7

2.00 2.00 2.17 2.17 2.00 2.00 2.02 2.02 1.92 2.13

0.80 : 0.20

0.26

7.34

0.85 : 0.15

0.19

7.45

0.84 : 0.16

0.16

7.36

0.88 : 0.12

0.16

7.46

0.91 : 0.09

0.12

7.55

This journal is © The Royal Society of Chemistry 2015

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Dalton Transactions

Fig. 7

Paper

Variation of TC with (a) IS/IM and (b) charge transfer energy (Δ) along the oxidation process. It shows that TC increases with decreasing Δ.

shows the variation of TC with Δ and IS/IM. It shows that Δ decreases with oxygenation in parallel with increasing TC. O 1s core level spectra. Fig. 8a shows the O 1s spectra for all samples. It is apparent that the O 1s spectra exhibit two broad peaks: a higher intensity one at ∼528.7 eV and another (shoulder) at ∼530.3 eV with lower intensity. However, both peaks are considerably broader for all the samples, which suggest the presence of overlapping multi-components in the samples. In the present system, the oxygen component is complex, since it includes three different sites: the SrO layers, the CuO2 planes and the (Mo/Cu)O1+δ site. The present O 1s emissions correspond to the emission from O ions bonded within three different metal–oxide planes: O ions bonded to Mo/Cu in (Mo/Cu)O1+δ planes, to Sr in SrO layers and to Cu in

the CuO2 plane. We have fitted the O 1s spectrum for the OA sample with three different components, and the fitted data are shown in Fig. 8b. From the fitting, we can assign the peak at 528.9 eV to O 2p–Cu 3d bonding, consistent with the studies of other cuprate superconductors,32 while the peak at about 530.3 eV is assigned to O 2p–Mo 3d bonding.11 The peak at 531.3 eV, on the other hand, is due to the O bonding to Sr in the SrO layer. As shown in Fig. 8a, the shoulder at 530.3 eV increases after oxygen annealing. The peaks at 528.9 eV and 530.3 eV correspond to the O ions bonded to the CuO2 planes and (Mo/Cu) O1+δ, respectively. So the enhancement of the shoulder at ∼530.3 eV has a correlation with the extra oxygen ions in the (Mo/Cu)O1+δ planes after oxygenation. The oxygenation enhances the amount of oxygen in the (Mo/Cu)O1+δ plane; this elevates the amount of the MoVI state and increases the number of Mo–O bonds and is reflected in the O 1s line shapes related to the O 2p–Mo 3d bonding. It is noteworthy that we have also observed significant changes in the Sr 3d core-level spectra after the introduction of extra oxygen ions in the (Mo/Cu)O1+δ plane. This suggests that the change in the Sr 3d core level spectra is correlated with the enhancement of the O 1s spectra at the higher energy side after oxygenation.

4.

Discussion

Surface electronic structure

Fig. 8 (a) Core level XPS spectra of the O 1s energy region for the Mo0.3Cu0.7Sr2ErCu2Oy samples. (b) Fitted O 1s spectrum for the OA sample. The peak at 528.9 eV is assigned to the O 2p–Cu 3d bonding, consistent with the studies of other cuprate superconductors,26 while the peak at about 530.3 eV is assigned to O 2p–Mo 3d bonding. The peak at 531.3 eV is due to the O bonding to Sr in the SrO layer.

This journal is © The Royal Society of Chemistry 2015

At the beginning of this discussion on the electronic states of the scraped Mo0.3Cu0.7Sr2ErCu2Oy samples, we emphasized that, in principle, we can only obtain average information from the surface. On the other hand, due to the scraping under ultrahigh vacuum (UHV; 10−9 mbar) at room temperature, depth sensitive information has been obtained which can be used to differentiate between surface and bulk electronic states. As we have seen, the Er 4d core level spectra do not show significant changes after oxygenation, as expected. However, the Sr 3d and O 1s core level spectra show a complicated

Dalton Trans., 2015, 44, 10795–10805 | 10801

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Paper

Dalton Transactions

scenario. As discussed earlier, two doublets and the changes in the Sr 3d core level spectra after oxygenation have two possible explanations. Nevertheless, the structural analysis12,13 using diffraction (NPD and XRD) patterns did not show any indication of intermixing of Sr and Er. In fact, the occupancies of Sr and Er become negative when they are allowed to share their opposite Er and Sr sites. At the same time, the Er 4d core level spectra do not show any significant change. Both of these observations rule out the possibility of the amalgamation of Sr and Er in the structure. Our detailed structural studies12,13 showed the accumulation of oxygen in the copper chain site ((Cu/Mo)O1+δ) after oxygenation: the amount of extra oxygen in the chain site (δ) changes from 0.34 ((Cu/Mo)O1.34 in AP) to 0.55 ((Cu/Mo)O1.55 in HPO), after HP oxygenation. As the Sr atom (SrO layer) resides between the (Cu/Mo)O1+δ chain and the CuO2 plane in the structure, the extra oxygen atoms in the copper chain ((Cu/Mo)O1+δ) do change the local chemical environment of the Sr atom, and this is reflected in the core level spectra of the Sr 3d. From the fitting of the O 1s spectrum, we have found that the O 1s emissions correspond to the O ions bonding in the three different metal–oxygen planes. It also shows that the shoulder around 530.3 eV corresponds to the O 2p–Mo 3d bonding that increases after oxygenation. Therefore, our study suggests that the observed changes in the Sr 3d and O 1s spectra are correlated. The extra oxygen after oxygenation increases the average number of oxygen bonds with Sr and Mo, which changes the chemical environment of the Sr atom as is reflected in the core level spectra of the Sr 3d and the O 1s. Redox process. Fig. 9 shows the relative energy shift of the Mo 3d and Cu 2p core level spectra with oxygenation. This emphasizes the shift of the Mo 3d and Cu 2p core level spectra in opposite sides of the energy region; this indicates the simultaneous oxidation of MoV and reduction of CuIII. Fitting the Mo 3d and Cu 2p core level spectra allows us to propose a very interesting doping mechanism upon oxidation: MoV x MoVI ð0:3xÞ þ CuIII y CuII ð2:7yÞ $ MoV ðxmÞ MoVI ð0:3xþmÞ þ CuIII ðynÞ CuII ð2:7yþnÞ

ð3Þ

where 0 ≤ (x − m) ≤ 0.3 and 0 ≤ (y − n) ≤ 2.7. This apparently contradictory redox mechanism is feasible because copper(III) oxides tend to decompose at moderate temperatures. KCuO2 decomposes40 under an oxygen atmosphere at 760 K, a tendency that is characteristic of the higher oxidation states of the elements of the first transition series; likewise, CrVIO3 decomposes below 570 K. However, in the second and third transition metal rows, the higher oxidation states are more stable. In the case of the oxides of Mo and W, in the same group (VI) as chromium, the highest oxidation state (VI) is the most stable: MoVIO3 sublimes in air without decomposing at temperatures above 1430 K.41 As we have copper on two different sites (Cu1 and Cu2), it is not possible to estimate the accurate hole concentration in the CuO2 plane (nor the oxidation state of copper for each individ-

10802 | Dalton Trans., 2015, 44, 10795–10805

Fig. 9 The relative energy shift of the Mo 3d and Cu 2p core level spectra with oxygenation in Mo0.3Cu0.7Sr2RECu2Oy. It emphasizes the shift of the Mo 3d and Cu 2p core level spectra in opposite sides of the energy region, pointing out the simultaneous oxidation of Mo and reduction of Cu.

ual site) from the present XPS results. Yet TEP measurements, a bulk property, show the enhancement of the hole concentration with increasing oxygen content. It then appears that the oxidation of MoV to MoVI, present only in the (Cu/Mo) chains, produces enough holes so as to compensate for the reduction of CuIII and TC, increases, a substantial amount, up to 84 K (HPO sample). Also, the sample (HPO_8 GPa) oxidized at 80 kbar shows the onset of a superconducting transition temperature at 89 K, the highest ever observed in this family. Although the present oxidation conditions, 50 kbar and 80 kbar at 773 K, are already highly oxidising, the ever-increasing trend in the functionality of TC with such parameters suggests that even higher TC values could be obtained under extreme conditions, since we have not attained the overdoped regime. Charge transfer energy and orbital admixture As shown in Fig. 7, the superconducting transition temperature shows an increasing trend with decreasing IS/IM and, consequently, Δ. Studying the Bi2Ca1−xRExSr2Cu2O8+s (RE = Rare Earth) series, Rao et al.42 showed that the satellite intensity goes through a minimum at around the same composition where the hole concentration and TC show maxima. But the oxygen contents were assumed to be the same for all compounds in their studies. However, a recent study with YBCO showed that the IS/IM ratio (and then the charge transfer energy Δ) increases with increasing oxygen content, and this is associated with an increase in TC.43 It is worth recalling here that our materials have a structure similar to YBCO, which prompted us to compare our results with those of YBCO. In Mo0.3Cu0.7Sr2ErCu2Oy we have observed a decrease in both the IS/IM ratio and the overall oxidation state of copper after oxy-

This journal is © The Royal Society of Chemistry 2015

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Dalton Transactions

Paper

Fig. 10 Variation of the superconducting transition temperatures (TC) with (a) the apical Cu2–O2 distances and (b) buckling angle (O1–Cu2–O1 angle) for different MoxCu1−xSr2RECu2Oy (RE = Y, Er and Tm, ref. 6, 7 and 10) samples. It clearly shows that TC increases in parallel with the shortening of the apical O distance (Cu2–O2) and increasing CuO2 plane buckling, while the two orbital model17,45 predicts the opposite effect (see the text).

genation (Table 4); this is correlated with the decrease in the charge-transfer energy. Besides the interesting evolution of the electronic states, some remarkable and unconventional structural points have also been observed for the present materials. As indicated above, in all the studied 1212-molybdo-cuprate compounds12,13,16 a contraction in the apical Cu2–O2 (dapical ) bond length is observed after oxygenation (Fig. 10a) with increasing TC. The role of apical oxygen (O2 in the present materials) in high-TC superconductivity has been suggested to have great importance.44–49 For instance, empirically, the presence of the apex oxygen atoms makes the CuO2 plane easier to dope with holes. The compounds with square coordination have not been doped successfully but those with copper in octahedral coordination have been over-doped.50 Also, the study of the effect of pressure on TC shows that the T*-phase superconductor with pyramidal copper coordination (having oxygen atoms at the apex) has a very large pressure enhancement of TC. On the other hand, in the T′-phase superconductor where copper is square planar coordinated, the pressure coefficient of TC is negligibly small.45 The structural results then clearly indicate that the apical oxygen indeed plays an important role in assisting charge transfer between the (Mo/Cu)O1+δ chain and the CuO2 copper plane in the transformation from semiconducting to superconducting. On the other hand, Sakakibara et al.51 proposed that shortening the apical O distance (Cu2–O2) increases the admixture of the Cu 3d3z2−r2 orbital with the inplane Cu 3dx2−y2 orbital, and this decreases TC. In fact, in single-layer cuprates, increasing the apical oxygen distance reduces the charge-transfer energy, which increases the superconducting transition temperature.51,52 We can recall that HgBa2Ca2Cu3O8+δ shows the largest TC having a large apical distance and an almost flat copper plane.25,51–54 Therefore, shortening the apical oxygen distance in the oxygenated superconducting (Mo,Cu)-1212 samples could worsen the hole

This journal is © The Royal Society of Chemistry 2015

transfer (charge transfer) from the copper chain site. Nevertheless, our TEP measurements show that the hole concentration in the CuO2 plane increases after oxygenation in the present materials. Recent theoretical work by Yee et al.55 in contrast to Sakakibara et al.51 proposes that the charge transfer energy controls the strength of correlations and therefore tunes the maximum superconducting transition temperature. Starting with the most correlated cuprate, La2CuO4, reducing the charge transfer energy enhances TC. Chmaissem et al.56 showed that the hole concentration per CuO2 plane would increase with the displacement of apical O toward the CuO2 plane after oxygenation. At the same time, moving the apical oxygen towards the copper plane makes the copper plane buckle, increasing electron–electron repulsion; this is supposed to damage the superconducting properties. However, it is worth emphasizing here that structural studies on the present and related materials12,13,16 have shown the enhancement of the buckling in the CuO2 plane after oxygenation in parallel with increasing TC (Fig. 10b). Therefore, in agreement with Yee et al.55 our experimental results indicate that charge transfer energy tunes the superconducting transition temperature (TC) and in the present system, TC increases with decreasing charge transfer energy. This occurs in parallel with the shortening of the apical O distance (Cu2–O2) and with the increasing CuO2 plane buckling, while the two orbital model51 predicts the opposite effect.

5. Conclusions The detailed studies on the electronic states for the Mo0.3Cu0.7Sr2ErCu2Oy samples and the influence of oxygenation on the same are presented here. The present results clearly corroborate the all-important role of apical oxygen in the strength of the charge transfer between the charge reser-

Dalton Trans., 2015, 44, 10795–10805 | 10803

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Paper

voir layer – (Mo0.3Cu0.7)O1+δ chains – and the CuO2 superconducting copper planes in the transformation from a semiconducting to a superconducting state; however, this relationship does not seem to have a single-type of functionality. The XPS studies on the Mo 3d and Cu 2p core level spectra allow us to propose a very interesting doping mechanism upon oxidation. The changes in the Sr 3d and O 1s core level spectra are related to oxygen insertion in the (Mo/Cu)O1+δ chain sites. Analysis of the Cu 2p core level spectra within the Configuration Interaction (CI) model suggests that higher values of TC are related to low values of the O 2p–Cu 3d charge transfer energy, which is in contrast to the accepted views, occurring in parallel with the shortening of the apical copper–oxygen distance (Cu2–O2) and the increasing of the CuO2 plane buckling angle. Under the present oxidizing conditions, TC values up to 89 K have been obtained. From the observed trend in the relationship between TC and PO2, one may expect even higher TC’s or, alternatively, a change from the underdoped to the overdoped regime.

Acknowledgements The authors are indebted to the European Union through the SOPRANO project (Seventh Framework Programme FP7/ 2007–2013 under Grant Agreement no. 214040) for funding this work and providing a Marie Curie grant to S. Marik. REE (Red eléctrica de España), the Spanish ‘Ministerio de Ciencia e Innovacion’ and ‘Comunidad de Madrid’ are also acknowledged for financial support given through projects REE/UCM 2014, MAT2010-19460 and S2009/PPQ-1626, respectively.

Notes and references 1 M. Murakami, Y. Ohishi, N. Hirao and K. Hirose, Nature, 2012, 485, 90–94. 2 O. Tschauner and C. Ma, Bridgmanite, 2014, IMA 2014017. CNMNC Newsletter No. 21, August 2014, page 798; Mineralogical Magazine, 78, 797–804. 3 R. H. Mitchell, in Perovskites: Modern and Ancient, Almaz, 2003. 4 J. G. Bednorz and K. A. Muller, Z. Phys. B: Condens. Matter, 1986, 64, 189–193. 5 J. P. Attfield, J. Mater. Chem., 2011, 21, 4756–4764. 6 R. Coontz, in Science’s Top 10 Breakthroughs of 2013, 19 December 2013. 7 J. G. Bednorz and K. A. Muller, Phys. Rev. B: Condens. Matter, 1986, 50, 3458. 8 M. Imada, A. Fujimori and Y. Tokura, Rev. Mod. Phys., 1998, 70, 1039. 9 L. Bauernfeind, W. Widder and H. F. Braun, Physica C, 1995, 254, 151–158. 10 P. H. Hor, R. L. Meng, Y. Q. Wang, L. Gao, Z. J. Huang, J. Bechtold, K. Forster and C. W. Chu, Phys. Rev. Lett., 1987, 58, 1891.

10804 | Dalton Trans., 2015, 44, 10795–10805

Dalton Transactions

11 S. Marik, E. Moran, C. Labrugere, O. Toulemonde and M. A. Alario-Franco, J. Solid State Chem., 2012, 191, 40–45. 12 S. Marik, A. J. Dos Santos-Garcia, E. Morán, O. Toulemonde and M. A. Alario-Franco, J. Phys.: Condens. Matter, 2013, 25, 165704. 13 S. Marik, A. J. Dos Santos-Garcia, E. Morán, O. Toulemonde and M. A. Alario-Franco, Supercond. Sci. Technol., 2015, 28, 045007. 14 Q. Xiong, Y. Y. Xue, J. W. Chu, Y. Y. Sun, Y. Q. Wang, P. H. Hor and C. W. Chu, Phys. Rev. B: Condens. Matter, 1993, 47, 11337. 15 V. P. S. Awanaa, A. Gupta, H. Kishan, E. Takayama-Muromachi, T. Watanabe, M. Karppinen, H. Yamauchi, S. K. Malik, W. B. Yelon, V. Ganesan and A. V. Narlikar, Solid State Commun., 2003, 129, 117–121. 16 I. Grigoraviciute, H. Yamauchi, M. Karppinen and M. Marezio, Phys. Rev. B: Condens. Matter, 2010, 82, 104507. 17 I. Grigoraviciute, M. Karppinen, C. Ting-Shan, L. Ru-Shi, C. Jin-Ming, O. Chmaissem and H. Yamauchi, J. Am. Chem. Soc., 2010, 132, 838–841. 18 H. L. Tsay, C. R. Shih, Y. C. Chen, W. H. Lee, T. H. Meen and H. D. Yang, Physica C, 1995, 252, 79–86. 19 H. L. Tsay, Y. C. Chen, S. S. Weng, C. F. Chang and H. D. Yang, Phys. Rev. B: Condens. Matter, 1999-I, 59, 636– 640. 20 S. F. Hu, R. S. Liu, S. C. Su, D. S. Shy and D. A. Jefferson, J. Solid State Chem., 1994, 112, 203. 21 R. L. Harlow, G. H. Kwei, R. Suryanarayanan and M. A. Subramanian, Physica C, 1986, 257, 125–136. 22 T. A. William, S. R. Harrison, J. T. Vaughey, L. Liu and A. J. Jacobson, J. Solid State Chem., 1995, 119, 115–119. 23 H. Sakakibara, H. Usui, K. Kuroki, R. Arita and H. Aoki, Phys. Rev. Lett., 2010, 105, 057003. 24 B. Buchner, M. Breuer, A. Freimuth and A. P. Kampf, Phys. Rev. Lett., 1993, 73, 1841–1844. 25 X. Zhang, W. H. Lu and C. K. Ong, Physica C, 1997, 289, 99–108. 26 F. U. Hillebrecht, J. Fraxedas, L. Ley, H. J. Trodahl, J. Zaanen, W. Braun, M. Mast, H. Petersen, M. Schaible, L. C. Bourne, P. Pinsukanjana and A. Zettl, Phys. Rev. B: Condens. Matter, 1989, 39, 236. 27 Y. Ando, A. N. Lavrov, S. Komiya, K. Segawa and X. F. Sun, Phys. Rev. Lett., 2001, 87, 017001. 28 J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, Handbook of X-Ray photoelectron spectroscopy, Physical Electronics, U.S.A., Inc., 1995. 29 M. Khaled, P. Srivastava, B. R. Sekhar, K. B. Garg, S. K. Agarwal, A. V. Nalikar and F. Studer, J. Phys. Chem. Solids, 1998, 59, 777–782. 30 Y. Uwamino, Y. Ishizuka and H. Yamatera, J. Electron Spectrosc. Relat. Phenom., 1984, 34, 67. 31 A. Katrib, V. Logie, M. Peter, P. Wehrer, L. Hilaire and G. Maire, J. Chim. Phys. Phys.-Chim. Biol., 1997, 94, 1923. 32 S. Larsson, Chem. Phys. Lett., 1976, 40, 362. 33 G. Van der Laan, C. Westra, C. Haas and G. A. Sawatzky, Phys. Rev. B: Condens. Matter, 1981, 23, 4369.

This journal is © The Royal Society of Chemistry 2015

View Article Online

Published on 20 February 2015. Downloaded by University of Michigan Library on 17/06/2015 07:32:04.

Dalton Transactions

34 F. Parmigiani and L. Sangaletti, J. Electron Spectrosc. Relat. Phenom., 1994, 66, 223. 35 R. P. Vasquez, D. L. Novikov, A. J. Freeman and M. P. Siegal, Phys. Rev. B: Condens. Matter, 1997-I, 55, 14623. 36 F. C. Zhang and T. M. Rice, Phys. Rev. B: Condens. Matter, 1988, 37, 3759. 37 J.-J. Yeh, L. Lindau, J.-Z. Sun, K. Char, N. Missert, A. Kapitulnik, T. H. Geballe and M. R. Beasley, Phys. Rev. B: Condens. Matter, 1990, 42, 8044. 38 K. Okada and A. Kotani, J. Phys. Soc. Jpn., 1989, 58, 2578– 2585. 39 K. Allan, A. Campion, J. Zhou and J. B. Goodenough, Phys. Rev. B: Condens. Matter, 1990, 41, 11572. 40 G. A. Costa and E. Kaiser, Thermochim. Acta, 1995, 20, 591– 598. 41 N. N. Greenwood and A. Earnshaw, Chemistry of the elements, Pergamon Press, Oxford, 1984. 42 C. N. R. Rao, G. Rao Ranga, M. K. Rajumon and D. D. Sarma, Phys. Rev. B: Condens. Matter, 1990, 42, 1026. 43 K. Maiti, J. Fink, S. Jong, M. Gorgoi, C. Lin, M. Raichle, V. Hinkov, M. Lambacher, A. Erb and M. Golden, Phys. Rev. B: Condens. Matter, 2009, 80, 165132. 44 E. Kaldis, P. Fischer, W. Hewat, E. A. Hewat, J. Karpinski and S. Rusiecki, Physica C, 1989, 159, 668.

This journal is © The Royal Society of Chemistry 2015

Paper

45 C. Murayama, N. Mori, S. Yomo, H. Takagi, S. Uchida and Y. Tokura, Nature, 1989, 339, 293. 46 K. A. Muller, Z. Phys. B: Condens. Matter, 1990, 80, 193. 47 H. Matsukawa and H. Fukuyama, J. Phys. Soc. Jpn., 1990, 59, 1723. 48 T. Tohyama and S. Maekawa, J. Phys. Soc. Jpn., 1990, 59, 1760; T. Tohyama and S. Maekawa, Physica B, 1990, 165–166, 1019. 49 Y. Ohta, T. Tohyama and S. Maekawa, Phys. Rev. B: Condens. Matter, 1991, 43, 2968. 50 J. B. Torrance, A. Bezinge, A. I. Nazzal, T. C. Huang, S. S. P. Parkin, D. T. Keane, S. J. LaPlaca, P. M. Horn and G. A. Held, Phys. Rev. B: Condens. Matter, 1989, 40, 8872. 51 H. Sakakibara, H. Usui, K. Kuroki, R. Arita and H. Aoki, Phys. Rev. Lett., 2012, 85, 064501. 52 C. Weber, C. Yee, K. Haule and G. Kotliar, EPL, 2012, 100, 37001. 53 O. Chmaissem, Q. Huang, E. V. Antipov, S. N. Putilin, M. Marezio, S. M. Loureiro, J. J. Capponi, J. L. Tholence and A. Santoro, Physica C, 1993, 217, 265. 54 C. W. Chu, L. Gao, F. Chen, Z. J. Huang, R. L. Meng and Y. Y. Xue, Nature, 1993, 365, 323–325. 55 C.-H. Yee and G. Kotliar, Phys. Rev. B: Condens. Matter, 2014, 89, 094517. 56 O. Chmaissem, J. D. Jorgensen, S. Short, A. Knizhnik, Y. Eckstein and H. Shaked, Nature, 1999, 397, 45.

Dalton Trans., 2015, 44, 10795–10805 | 10805